Assessment of performance-driven investment strategies of distribution systems using reference networks V. Levi, G. Strbac and R. Allan Abstract: Distribution systems are inherent monopolies and therefore they have generally been regulated to protect customers and to ensure cost-effective operation. In the UK this is one of the functions of the Office of Gas and Electricity Markets (OFGEM). Initially the regulation was based on the value of assets but there is now a trend towards performance-based regulation. To achieve this, a methodology is needed that enables the reliability performance associated with alternative investment strategies to be compared with the investment cost of these strategies. At present there is no accepted approach for such assessments. An approach based on the concept of reference networks is proposed. The essential idea is to analyse a small number of networks that represent the real feeders of the system being considered. The main stages are: to classify the real feeders into coherent clusters, to construct a representative network for each cluster, to assess the reliability of these representative networks and to re-aggregate the results to give those for the overall system. These representative networks are then analysed repetitively for different investment scenarios and the corresponding reference networks derived. The approach can be used to assess individual companies on an absolute basis or to compare the relative performance between companies. The approach, its implementation in conjunction with a number of distribution companies, and its effectiveness when approach using the underground part of a real system are described. 1 Introduction Since the privatization of the electricity supply industry (ESI) in the UK, the regulator (OFGEM) has been responsible for controlling the revenue recovered by distribution network operators (DNOs) to ensure cost- effective operation and to protect customers. Initially, the regulation was based on the value of assets owned by the DNOs. Although this is still predominantly the case, considerable interest has been shown worldwide in moving from this asset-based regulation to performance-based regulation [1–6] , to relate capital and operational expendi- ture to the benefit that customers derive from such expenditure. This implies that the overall distribution revenue is a function not only of the capital and operational costs incurred by a DNO, but also of the quality of service they provide to customers. In the UK, limited aspects of performance-based regulation have been introduced through the information and incentives project (IIP) [7] . The primary objective of this IIP is to establish a framework in which the quality of supply will be properly valued and incentivised. In the initial IIP proposal [7], the output measures upon which financial incentives should be applied are the number of interruptions, their duration, and the telephone response that customers receive. Of these indicators, the greatest weight is given to the continuity of supply indices, placing the reliability performance of a company and the sub- sequent cost-benefit analysis at the centre of attention. However, a major issue to be resolved is related to the exact implementation of the incentive schemes [8]; one key dilemma being whether to base the incentive mechanism on ‘absolute’ or ‘relative’ performance [8] . Under the former approach, adjustments to the revenue would be based on assessing the company’s performance against its own individual performance target. This means no explicit comparison with other companies’ performance, and market forces would not drive companies’ investment strategies and operation policies [9, 10]. On the other hand, under the relative performance approach, adjustments to the revenue would be assessed by comparing each company’s performance against all other DNOs using benchmarks. The proposal by OFGEM [11] makes use of the ‘relative’ approach, in which comparison, achieved by a normalisation process, between companies’ performances is done with the aid of the average performance of all companies. The major concerns with this approach are the adequacy of using the average performance as a comparison benchmark, and the difficulties associated with comparing companies supplying different customer densities and having very diverse performances due to operating very different types of networks. This paper proposes an approach, based on the concept [12] of a ‘reference network’, which overcomes all the difficulties associated with the relative approach. Although primarily developed to provide an absolute assessment of the performance and for analysis of alternative investment scenarios, it can also be used objectively for comparing the The authors are with the Ferranti Building B7, Manchester M60 1QD, UK University of Manchester E-mail: goran.strbac@manchester.ac.uk r IEE, 2004 IEE Proceedings online no. 20041109 doi:10.1049/ip-gtd:20041109 Paper first received 2nd September 2003 and in revised form 20th July 2004 IEE Proc.-Gener. Transm. Distrib., Vol. 152, No. 1, January 2005 1